[ARGUS] Segmentation Fault with 2.0.6rc2 on FreeBSD 4.9-RELEASE

[Peter Van Epp]

On Fri, Apr 16, 2004 at 12:15:27AM -0500, eric wrote:
> On Thu, 2004-04-15 at 21:54:29 -0700, Peter Van Epp proclaimed...
> 
> > 	Can't speak for Eric, but all of my boxes pretty much (work and home)
> > run ECC ram. Its not worth my time to fight memory problems when a small 
> > capital outlay for ECC eliminates it as a possibility. From the traces Eric
> > provided earlier it looks like something in the code is 0ing a little far 
> > sometimes :-)
> 
> I've heard all this talk of bad RAM for years now; personally, all I
> use is *BSD type systems, and I've never seen an actual memory
> problem -- especially with a program termininating in the way that
> argus seems to be dying off in this manner.
> 
> I've been restarting argus often now and it seems to be working well
> -- I know this is a hack, but....well....sometimes they seem to
> work! :)
> 
> - Eric

	The 20 year old 386 that runs argus on my DSL line at home has memory
problems sometimes, but it also has parity (if not ECC) ram and reacts with an
NMI rather than corruption. In that machine's hayday OS2 was the preferred
memory test application, it used enough of ram in odd ways that it would often
expose problem memory (and even parity was reasonably unusal in those days).
I haven't even seen an ECC hit in years on any of our machines so I suspect 
by and large memory problems are a thing of the past these days. 

Peter Van Epp / Operations and Technical Support 
Simon Fraser University, Burnaby, B.C. Canada